Abstract
Semantic priming of morphologically complex Dutch verbs was investigated in two cross-modal experiments. Spoken sentences served as primes to visual target words, presented for lexical decision. Targets were either simple verbs (e.g., brengen, bring) or particle verbs, consisting of a simple verb plus a separable prefix (e.g., meebrengen, bring along). Particle verbs were either semantically transparent or opaque. Transparent particle verbs are semantically related to their constituent verb (e.g., meebrengen, bring along), opaque verbs are not (e.g., ombrengen, exterminate). In Experiment 1, facilitation was consistently obtained when verb targets were semantically congruent with the content of the sentence primes. But priming was also found in incongruent conditions, when opaque verbs served as targets with sentences constructed to prime the meaning of their embedded verbs. Post-hoc analyses and data from Experiment 2, however, showed that this was due to the ambiguous nature of some opaque particle verbs. Whereas the dominant opaque and the subordinate transparent meaning of ambiguous particle verbs could both be primed, truly opaque verbs were not facilitated by the semantic field of their constituent verbs. Given abundant evidence for a close association, at a morphological level, of transparent and opaque complex words to their constituents, the data demonstrate a dissociation between connections at lexical and conceptual levels of representation.
Acknowledgments
All experiments were carried out in the Netherlands, at the Max Planck Institute for Psycholinguistics. We thank the Institute and the members of the technical group for their support. We are grateful to Ram Frost and two anonymous reviewers, for their valuable comments. This research was supported by a grant (Dr. 229/2) from the Deutsche Forschungsgemeinschaft (DFG) to Drews and Zwitserlood.
Notes
1We use the term ‘simple verb’ throughout to refer to the combination of a root (e.g., plakk-) and the infinitival inflection (-en). Simple verbs contained in particle verbs are also referred to as ‘constituent verbs’ or ‘embedded verbs’.
2We use semantic and conceptual interchangeably and consider this information to be outside of, but closely linked to, the language system proper.
3Word frequency is difficult to assess for particle as well as for simple verbs. In the CELEX database for Dutch (see Baayen, Piepenbrock, & van Rijn, 1993), every occurrence of a verb in the text corpus adds to the lemma of the simple verb. This implies that particle verbs in finite form add to the count of their constituent simple verbs, leading to an overestimation of their frequency and an underestimation of the frequency of particle verbs. The frequency count of particle verbs is based only on those cases in which the verb was written as one word.
4The clear dominance of the opaque meaning of these verbs is evident from the semantic pretest data and from the fact that none of us had ever reckoned with a secondary transparent sense before we started close-reading the dictionary.